• information theory
• consciousness
• computational neuroscience

Sleep is the most familiar alteration of consciousness. A recent paper showed that during non-REM sleep there is a breakdown of cortical connectivity, which might be involved in the fading of consciousness. By employing a large scale model of the thalamocortical system we investigated the differences in brain connectivity at a microscopic scale between wakefulness and sleep. Our work consisted of three phases: setting the model parameters in order to get realistic waking and sleep, comparing our simulations to real data and, finally, analyzing our data with an information theoretical approach. Our results show that: during wakefulness i) single conscious states last 100-150 ms, ii) minicolumns are more effective than neurons and columns in transferring information, iii) groups of nearby minicolumns have a high degree of connectivity, iv) during sleep there is a breakdown of connectivity between minicolumns, due to both the oscillatory pattern of activity (slow wave activity) and a stereotyped way of firing.